Article pubs.acs.org/Biomac
Wood Hydrolysate Barriers: Performance Controlled via Selective Recovery Anas Ibn Yaich, Ulrica Edlund, and Ann-Christine Albertsson* Fiber and Polymer Technology, Royal Institute of Technology (KTH), SE-100 44 Stockholm, Sweden S Supporting Information *
ABSTRACT: Films and coatings were produced from a noncellulosic polysaccharide-rich wood hydrolysate (WH), and the resulting oxygen barrier performance was improved by a selective choice of upgrading conditions. The WH was obtained from process water in the hydrothermal treatment of hardwood and subjected to one of three alternative upgrading treatments, resulting in xylan-rich fractions with significant differences in structure, composition, and properties of the recovered WH fractions, which in turn had a major impact on their performance with respect to tensile and oxygen barrier properties. The WH in the least upgraded state, the crudest fraction, produced films with the best performance in terms of oxygen permeability and was superior to corresponding films based on highly purified hemicellulose.
■
INTRODUCTION The biorefinery concept is an old approach with renewed interest in recent time, representing all the processes that separate renewable resources into different fractions that can be further utilized in various applications.1 Within this context, the industrial recovery, isolation, and utilization of noncellulosic fractions from wood is attracting ever increasing interest. Considerable amounts of hemicelluloses together with lignin fragments, pectins, and various low molecular weight compounds are released into the process water in many forestry processes, for example, thermomechanical pulping,2 fiber-board production,3 and in the prehydrolysis step that precede Kraft pulping,4 making forestry industry process water fractions, wood hydrolysates (WH), stand out a low-cost sources of great potential for the large scale conversion into, and production of natural materials, bulk chemicals and fuels. A range of methods for recovery and purification have been proposed to upgrade these forestry process waters and to extract soluble and dispersed compounds. For instance, the hemicellulose, O-acetylated galactoglucomannan, the major hemicellulose in softwoods, such as spruce and pine, has been recovered from thermomechanical pulp and fiber-board production process waters by combining different techniques such as microfiltration, ultrafiltration, diafiltration, ethanol precipitation, and flocculation.3,5−8 The choice of recovery routes and parameters has a large impact on the resulting product composition, molecular weights, and purity. The complete separation and isolation of wood constituents is, however, not straightforward due to the strong interactions between these different components. Actually, it has been suggested that the hemicelluloses−lignin interactions comprise covalent linkages in the form of lignin−carbohydrates complexes, LCCs,9 which is believed to make their separation © 2011 American Chemical Society
even more difficult. Consequently, the production of highly purified hemicelluloses typically requires tedious and costly multistep purification procedures that involve chemical demanding extractions and delignification, as well as chromatography techniques.8 Traditionally, the development of hemicellulose-based products has focused on the utilization of highly purified hemicelluloses,5,10−14 but the high production cost is a drawback from a commercial production perspective and limits the use of hemicelluloses to very high value-added applications. The use of less-purified, hemicellulose-rich WH is economically much more feasible and, in addition, it was recently shown that hydrolysate-based products such as hydrogels,6 barrier films, and coatings offer similar or even superior performances in comparison to those made from highly purified hemicelluloses.15−17 Utilization of noncellulosic polysaccharide-rich WH rather than purified hemicelluloses is more straightforward and enables the preservation of some of the native component interactions adding to the macroscopic performance. On one hand, the great variety of WH structures and performances possible to produce by varying the origin and preparation presents many advantages, not in the least in terms of versatility and the possibility of tailor-making a renewable resource for a given product behavior. On the other hand, the selection of a well-functioning WH for any given application is far from trivial because the WH structure and composition (and, hence, the properties) will vary over a broad range, depending on the recovery route. Received: October 27, 2011 Revised: December 15, 2011 Published: December 19, 2011 466
dx.doi.org/10.1021/bm201518d | Biomacromolecules 2012, 13, 466−473
Biomacromolecules
Article
Figure 1. Experimental scheme for upgrading of wood hydrolysate. respectively. All fractions obtained by centrifugation, ultrafiltration, and diafiltration were lyophilized and stored dry. Ultrafiltration. A volume of 2.5 L of WH was subjected to membrane filtration, in this case ultrafiltration, employing a tangential flow filtration cartridge unit equipped with a hydrophilic membrane made from regenerated cellulose (PLAC Prepscale, Millipore) with a nominal cutoff of 1000 Da and a surface of 100 cm2. The membrane filtration produced 0.5 L of a retentate, herein denoted WH1 (a high molecular weight fraction), and 2 L of a low molecular weight fraction permeate, LMW1. Diafiltration. A volume of 0.2 L of the high molecular weight fraction WH1 was fractionated further by diluting it with water to 1 L and then subjecting it to another membrane filtration down to a volume of 0.2 L of retentate WH2 and at the same time producing 1 L of a permeate LMW2. The same conditions as in the ultrafiltration step were employed. Ethanol Precipitation. One volume (50 mL) of the WH was mixed with nine volumes of ethanol (450 mL), stirred for 30 min to give a cloudy voluminous precipitate, and then left standing cold (8 °C) overnight. The voluminous precipitate denoted WH3 was recovered by centrifugation and decantation and then dried at 80 °C. The dissolved material retained in the supernatant LMW3 contained was recovered in solid form on a rotavapor at 70 °C. Composition and Structural Analysis. The identification and determination of the content of neutral anhydro monosugars in each WH fraction was performed by ion-exchange chromatography. Prior to the analysis, these samples were acid-hydrolyzed using a two-step technique according to TAPPI T222 om-83. The first step was at 30 °C for 1 h with 72% w/w H2SO4 and the second one was conducted at 120 °C for 1 h with 3% w/w H2SO4. The hydrolyzed solutions were then filtered through a glass fiber filter to remove the high condensate lignin (Klason lignin) from the monomeric sugar solution. Each filtrate was then diluted and analyzed by IC using high performance anion exchange chromatography (HPAEC-PAD, Dionex ICS-3000) equipped with a gradient pump, a pulsed amperometric detector (Dionex PAD) and a CarboPac PA1 separation column. NaOH/acetate gradient and water were used as eluents. Before each injection, the column was conditioned with 60% v/v of 300 mM NaOH (eluent B), and 40% v/v of 200 mM NaOH−170 mM NaC2H3O2 (eluent C) for 7 min and re-equilibrated with H2O (eluent A) for 7 min. The separation of the neutral monosaccharides was done by an isocratic step of H2O (eluent A) for 25 min. The flow rate of eluent and column
Our aim is to produce WH-based films and coatings where the oxygen barrier properties are improved through a selective design of upgrading routes in the recovery of the WH resource. Our hypothesis is that by controlling the recovery conditions, a fraction with a favorable combination of WH structure and composition is derived, resulting in a component mixture, allowing for strong interactions and, hence, the formation of a densely packed structure with lower oxygen permeability. Herein, a crude hardwood derived WH was upgraded by any of three alternative routes, ultrafiltration, ultrafiltration + diafiltration, or ethanol precipitation, and the fractions, respectively, were used to prepare thin films and coatings.
■
EXPERIMENTAL SECTION
Materials. The hardwood hydrolysate used in this work was kindly provided by Södra Cell AB, Sweden. It was produced from birch mixed with a small fraction of aspen wood chips by hydrothermal treatment in a batch reactor. The wood chips were steamed for 45 min at 100 °C, and then hot water was added until a liquid-to-wood ratio of 6:1 (volume/weight) was reached. The temperature of the mixture was ramped up to 160 °C at a rate of 1 °C/min and maintained at 160 °C for 1 h. The process water fraction was removed from the mixture and this WH was centrifuged for 2 min at 4000 rpm to remove insoluble residues. The centrifuged WH, having a pH of 4, was then stored at 4 °C until upgrading according to the routes shown in Figure 1 and explained in the next section. Highly purified xylan (BrX) from birch wood (CAS Number: 9014−63−5) and carboxymethyl cellulose (CMC) sodium salt with a medium viscosity of 400−1000 mPa·s, 2% in H2O (25 °C), and having a degree of substitution of 0.6−0.9 (CAS Number: 9004−32−4) were used as received from Sigma-Aldrich. Commercially used poly(ethylene terephthalate) (PET) films with a thickness of 38 μm were kindly supplied by Tetra Pak Packaging Solutions AB, Sweden. Upgrading Pretreatments of WH. The crude WH obtained from Södra cell AB was centrifuged and denoted WH0. Three different protocols were elaborated and used for the WH upgrading: ultrafiltration, ultrafiltration + diafiltration, or ethanol precipitation. Each method separated the WH into one high and one low molecular weight fraction, producing a total of six different hydrolysate fractions, 467
dx.doi.org/10.1021/bm201518d | Biomacromolecules 2012, 13, 466−473
Biomacromolecules
Article
temperature were constantly kept at 1.0 mL/min and 30 °C respectively. To facilitate the pH sensitive oxidation of monosaccharides, 300 mM NaOH was used as the postcolumn eluent stream. Five calibration standards with different concentrations were prepared using six monosaccharides: xylose, galactose, rhamnose, mannose, arabinose, and glucose. The Klason lignin residue was subsequently dried at 105 °C overnight and gravimetrically determined. The acid-soluble lignin was determined according to the NREL method.18 1 H NMR. The acetyl and 4-O-methyl-glucuronic acid contents were determined by 1H NMR. Samples (approximately 15 mg) were dissolved in 0.7 mL of DMSO-d6 in NMR tubes with an outer diameter of 5 mm. Spectra were recorded at room temperature at 400 MHz on a Bruker DMX-400 NMR spectrometer with Bruker software. The weight fraction of 4-O-methyl glucuronic acid W%(MeGlcpA) was calculated from the relative response of the 4-O-methyl glucuronic acid proton to the integrals of a known quantity of xylose according to eq 1:
W%(MeGlcpA) =
an inner diameter of 8.7 cm (area of 60 cm2) and left to dry at 50% relative humidity (RH) and at 23 °C until constant weight was reached (approximately 3 days). Coating Preparation. Wood hydrolysates WH0−WH3/CMC blends were prepared as previously described for film preparation maintaining the same WH/CMC concentrations. The mixtures were painted by hand onto 12 × 23 cm2 PET films using a round paint brush having bristle length of 3 cm and a diameter 1.5 cm. The coated films were allowed to dry at 50% RH and 23 °C for 2 days. Tensile Testing. The films were cut into rectangular specimens according to the ASTM D 638 standard,20 and conditioned for one week at 50% RH. For each film, a minimum of five specimens were tested to break using an Instron universal materials testing machine equipped with a 50 N load cell operating at a rate of 5 mm/min and with an initial grippe distance of 25 mm. Oxygen Permeability and Oxygen Transmission Rate. Oxygen permeability (OP) and oxygen transmission rate (OTR) of coatings were determined according to ASTM standard D3985-0221 using a Mocon Oxtran 2/20 (Modern Controls, Minneapolis, MN) equipped with a coulometric sensor. Samples were conditioned at 50% RH and 23 °C for 3 days. The sample area was 50 cm2, whereas the thickness was measured with a Mitutoyo micrometer as the mean value of five measurements. Hansen's Solubility Parameter Calculation. The Hansen solubility parameters (HSPs) approach is a widely used tool to predict the polymers solubility and estimate their compatibility. In this approach, the total cohesive energy density (δ) of a substance is considered to be the result of three contributing partial cohesive parameters; the nonpolar (δD), the dipole−dipole (δP), and the hydrogen bonding (δH) interactions as shown in eq 3.
I(MeGlcpA) M(MeGlcpA) × I(Xyl) M(Xyl) × W%(Xyl)
(1)
where I(MeGlcpA) and I(Xyl) are the integral of the signals ascribed to MeGlcpA H1 at 5.26 ppm and the sum of integrals ascribed to Xylose H1 at 4.40−4.49 ppm, respectively.19 M(MeGlcpA) and M(Xyl) are the respective molar mass values of the anhydro 4-Omethyl glucuronic acid (191 g/mol) and anhydro xylose (132 g/mol), and W%(Xyl) is the weight fraction of xylose that was determined based on the Ionic chromatography measurement. The degree of acetylation DSAc was calculated according to eq 2:
I(Ac)/3 DSAc = I(carbohydrate)/6
δ = (δ2D + δ2P + δ2H)1/2
(3)
These three parameters serve also as coordinates in a threedimensional system called Hansen's space. Each substance is represented as a point in the Hansen's space through its characteristic HSPs, which defines the center of its interaction sphere. The interaction sphere, having a radius R0, represents the boundaries of deviation allowed. The distance Ra between substance (a) and substance (b) is calculated by eq 4. In an ideal case, when (a) and (b) present a high mutual affinity, Ra < R0.
(2)
where I(Ac) and I(carbohydrate), respectively, are the relative integrals of the response of the acetyl groups at 2.00 ppm and carbohydrate at 3.2−5.6 ppm. It is assumed that all free acetic acid was eliminated during the lyophilization step of WH0−WH2 and during ethanol precipitation and drying for WH3. Theromogravimetric Analysis, TGA. The moisture and ash contents of the WH fractions were assessed with TGA. Around 5 mg of each sample was put into ceramic cups and loaded into a Mettler Toledo TGA/STDA 851e. The samples were heated from 30−700 °C at a rate of 10 °C/min under an air flow of 50 mL/min. Size Exclusion Chromatography, SEC. Molecular weights were determined with a Shimadzu SEC equipped with a refractive index detector and four PLgel 20 μm Mixed-A (300 × 7.5 mm) columns with an injection volume of 200 μL. Dimethyl acetamide with 0.05% w/w LiCl was used as the mobile phase (0.5 mL/min). The measurements were performed at 80 °C. The SEC system was calibrated with Pullulan standards with molecular weights ranging from 320 to 800000 g/mol. Before injection, samples were dissolved for 24 h at room temperature in dimethyl acetamide (DMAc, 0.05% w/w LiCl) and filtered (0.45 μm, Millipore). LC Solution software from Shimatzu was used for data acquisition and calculations. Film Preparation. CMC was selected as a cocomponent in WHbased film formulations following the encouraging results of previous work,15 showing that films comprised of a softwood hydrolysate and 40% w/w CMC had very low OP values and good mechanical properties. Films consisting of blends of either of the wood hydrolysates WH0−WH3 together with CMC were prepared by casting from water solution. A highly pure commercial xylan/CMC film was also prepared for comparison. A typical film was prepared according to the following protocol: 0.48 g of a WH or xylan and 0.32 g CMC were dissolved separately in 15 mL of deionized water by mixing on a shaking board at a shaking rate of 200 min−1 for at least 3 h. Once the components were completely dissolved, the CMC solution was added to the WH or the xylan solution. The mixture was stirred for another 12 h and poured into a polystyrene Petri dish with
Ra = ⎡4( δ − δ )2 + ( δ − δ )2 + ( δ − δ )2 )⎤1/2 Pa Pb Ha Hb ⎦ ⎣ Da Db
(4)
In this work, HSPs are used to assess the degree of affinity and compatibility first between the WHs constituents and also between the WHs and CMC cocomponent. First, the group contribution theory was used to calculate the HSPs δD, δP, δH, and δ of the repeating unit for xylan, lignin, and CMC according to their determined structures. For the CMC, an average carboxymethyl degree from 0.65 to 0.9 was used for the calculation. Then, HSPs were calculated for wood hydrolysates WH0−WH3 based on their respective compositions. Finally, the correlation between the different components was calculated using eq 4.
■
RESULTS AND DISCUSSION A WH was obtained from the hydrothermal treatment of hardwood and subjected to different recovery routes, as outlined in Figure 1. Ultrafiltration, diafiltration, and ethanol precipitation are considered to be the most adequate techniques for upgrading process waters or other fractions generated in the processing of lignocellulosic feedstock. The macromolecular fractions generated may be used as a green resource for production of films and coatings with high oxygen-barrier performance. Such a performance requires a good control of the chemical composition and structure of the resource. 468
dx.doi.org/10.1021/bm201518d | Biomacromolecules 2012, 13, 466−473
Biomacromolecules
Article
Table 1. Summary of WH Propertiesa BrX part 1
chemical composition (wt %)
molecular weight
part 2
content
part 3
yield tensile properties
arabinose rhamnose galactose glucose xylose total anhydro-sugar 4-O-methyl glucuronic acid acetyl Klason lignin acid-soluble lignin total lignin water ash Mn Mw PDI concentration of dry matter in the WH solutions (g/L) yield of WH (%)c tensile strength (MPa) tensile strain (%) E-modulus (GPa)
part 4
barrier properties
OTR (cm3 /m2 day) OP (cm3μm/m2 day kPa)
LMW1
LMW2
LMW3
n.d.e n.d.e n.d.e n.d.e ≥90b ≥90b b.l.d b.l.d n.d.e n.d.e n.d.e n.d.e n.d.e 2607 13835 5.3 n.d.e
1.11 0.28 b.l. 0.59 47.14 49.12 30.59 5.94 5.87 0.18 6.05 2.12 6.19 637 2102 3.29 59
WH0
b.l. b.l. b.l. b.l. 53.41 53.41 28.84 6.21 5.46 0.02 5.48 2.19 3.87 886 2651 2.99 117.25
WH1
b.l. b.l. b.l. b.l. 57.27 57.27 22.77 6.18 5.35 0.04 5.39 1.57 6.83 1187 3148 2.65 71.5
WH2
b.l. b.l. b.l. b.l. 58.60 58.60 24.23 4.91 0.78 0.05 0.83 3.62 7.81 1574 3849 2.44 29.9
WH3
8.31 2.45 1.78 5.29 36.32 36.32 b.l. 2.40 2.20 0.03 2.23 1.23 40.00 303 540 1.78 n.d.e
3.55 1.05 b.l. 1.67 36.93 43.2 19.65 2.20 5.80 0.03 5.83 3.28 25.85 408 737 1.8 n.d.e
3.90 0.96 b.l. 1.67 30.25 36.78 22.80 2.38 7.45 0.02 7.47 0.78 29.78 409 788 1.93 n.d.e
n.d.e 22.0 (±3.6) 3.9 (±0.7) 1.3 (±0.2) 8.1 (±0.5) 3.2 (±0.2)
99 19.0 (±15.0) 1.3 (±0.5) 1.6 (±0.1) 3.2 (±0.2) 1.3 (±0.1)
39.3 40.9 (±12.0) 1.5 (±0.4) 2.8 (±0.1) 10.5 (±0.4) 4.4 (±0.1)
23.98 48.6 (±9.7) 2.2 (±0.5) 2.5 (±0.2) 14.6 (±1.0) 6.19 (±0.31)
50.16 39.2 (±8.4) 1.3 (±0.1) 3.2 (±0.8) 14.6 (±1.55) 6.08 (±0.05)
n.d.e n.d.e
n.d.e n.d.e
n.d.e n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
n.d.e
a
Part 1: Chemical composition, average molecular weights, and polydispersity index (PDI) for commercial birch wood xylan BrX, untreated and upgraded wood hydrolysates WH0−WH3, and their respective low molecular weight fractions LMW1−LMW3. Part 2: concentrations and yields of dry matter for WH0−WH3. Part 3: Tensile properties of commercial birch xylan BrX and wood hydrolysates WH0−WH3 films containing 40% w/w of CMC. Part 4: oxygen transmission rate (OTR) and oxygen permeability (OP) of uncoated and coated PET films with highly purified commercial birch xylan (BrX) or wood hydrolysates WH0−WH3. bKnown chemical composition obtained from the supplier (Sigma Aldrich). cYield of WH [% w/w] = weight of the dry matter in each WH solution/weight of the dry matter in the crude process water ×100. db.l.: below the detection limit. e n.d.: not determined.
kPa−1. PET films coated with CMC showed an OP in the range of 12 cm3 μm m−2 day−1 kPa−1.24 Noteworthy, the crudest fraction (WH0) presented the best barrier and was significantly superior to the corresponding film prepared from highly purified Birch xylan (BrX). SEM images of WH-CMC-coated PET films show that WH-CMC coatings form continuous as well as macroscopically and microscopically homogeneous layers (Supporting Information, Figure S2). The mechanical properties from the uniaxial tensile tests are summarized in Table 1. Pure CMC films have an E-modulus of 3,2 GPa, a tensile strength of 77 MPa, and a strain-at-break of 8%. The films from upgraded WHs had higher E-modulus and tensile strength than those made from highly purified BrX and untreated WH. Particularly for the films made from WH0, the high value of standard deviation is due to the relative brittleness of these films that might lead to the introduction of microcracks during specimen preparation. Furthermore, the films made from up-graded WH had higher tensile strengths and E-moduli than the films made from BrX. Despite the small increase in tensile strain for the WH2based film, WH-based films have generally low tensile strains. Therefore, the WH material is suitable to be applied as barrier layer in multilayer package where the structural function of the package is ensured by the substrate.
Hydrolysates Performance. In the food packaging industry sector, low oxygen permeability (OP) is generally an important requirement for a packaging material to fulfill its function in preserving food products. Common food packaging films should have an OP lower than 38.9 cm3 μm m−2 day−1 kPa−1 in order to be regarded as good barriers.22 Hemicelluloses have previously been acknowledged as having great potential as barriers.23 For instance, films based on arabinoxylan and plasticized with 40% w/w sorbitol or glycerol had OPs of 4.7 and 7.4 cm3 μm m−2 day−1 kPa−1.12 Films made from purified O-acetylated galactoglucomannan (AcGGM) blended with 35% w/w CMC showed an OP as low as 1.28 cm3 μm m−2 day−1 kPa−1.5 Recently, it was shown that the partially upgraded softwood hydrolysate was superior to purified AcGGM and gave rise to films with an OP as low as 0.3 cm3 μm m−2 day−1 kPa−1.15 Herein, PET films coated with thin layers (∼4 μm in thickness) of WH0−WH3/CMC blends were explored with respect to oxygen barrier performance. As shown in Table 1, the oxygen barrier performance was found to improve significantly for all the WH/CMC-coated films as compared to the uncoated PET film. The highest oxygen barrier effect was observed for the coating made from WH0, which gives an OP of 1.3 cm3 μm m−2 day−1 kPa−1, while the highest permeability was observed for WH3, having an OP of 6.1 cm3 μm m−2 day−1 469
dx.doi.org/10.1021/bm201518d | Biomacromolecules 2012, 13, 466−473
Biomacromolecules
Article
Figure 2. Typical structural motif of O-acetyl-4-O-methylglucuronoxylan.
Figure 3. 1H NMR spectra (left) of the wood hydrolysates WH0−WH3 and the highly purified birch xylan BrX. 1H NMR spectrum (right) of WH0 with the region 5.4−4.3 ppm magnified as an inset.
structures and compositions, affected by the upgrading parameters. The chemical composition of each one of wood hydrolysates (WH0−WH3) and their corresponding low molecular weight fractions (LMW1−LMW3) are shown in Table 1. It can be observed that xylan hemicellulose (determined as the sum of xylose and MeGlcpA) is the major constituent of the untreated WH, although significant amounts of lignin (determined as the sum of Klason and acid soluble lignins), ash, water, and traces of pectin (arabinose + glucose + rhamnose + galactose) were also detected. Xylan tends to be retained in the high molecular weight (HMW) fractions during upgrading, while pectin was preferably released to the low molecular weight (LMW) fractions. The xylan content in the WH was increased during all three investigated upgrading procedures, reaching the highest enrichment of xylan in the case of ultrafiltration (WH1), indicating that, during the diafiltration step, some xylan was not retained in the WH. The lignin content, including the Klason and acid soluble lignins, was reduced from 6% w/w to 1% w/w by ethanol precipitation, but it remained constant after ultrafiltration and diafiltration. The retention of lignin after membrane filtration can be explained by its high molecular weight and its hydrophobic character which would increase it retention in the retentate by the membrane. The lignin residue gives rise to
Obviously, the WH performance in terms of mechanical and oxygen barrier properties differs to a large extent depending on the adopted recovery route. These differences between the WH fractions are attributed to several possible factors. Changes in the chemical composition, structure, and molar mass distribution induced by the different upgrading methods constitute the major ones. The same property−composition− structure relationships that exist in similar systems such as natural plants enable us also to achieve the desired properties of the WH material. In wood, for instance, the amounts of each macromolecular constituent, that is, cellulose, hemicellulose, and lignin, are optimized to ensure perfectly its mechanical and biological functions. Likewise, the chemical structures of wood components are also well adapted to maximize their mutual interactions. In hardwood, such as birch, the hemicelluloses consist mainly of O-acetyl-4-O-methylglucuronoxylan.19,25 A typical segment of this molecule is shown in Figure 2. The Oacetyl-4-O-methylglucuronoxylan chain segments polarity is tailored according to their local environment by adjusting the amount and distribution of acetyl and 4-O-methylglucuronic acid side groups along the chain, thereby increasing the overall compatibility and adhesion of the woody material.26 Hydrolysates Composition. The obtained differences in the WH performance can be related to their individual 470
dx.doi.org/10.1021/bm201518d | Biomacromolecules 2012, 13, 466−473
Biomacromolecules
Article
significantly according to the adopted recovery and upgrading routes that enable us to tune the WH mechanical and barrier performances. Noteworthy, the WH subjected to the least extensive upgrading showed the best performance in terms of OP. Furthermore, the increase in average molecular weights observed after upgrading pretreatments of the WH leads on one hand to a decrease in the oxygen barrier function of the coating, but on the other hand to an enhancement in the mechanical properties of the films, as shown by comparing films made from WH1 and WH2, having the same lignin content of (6 wt %) and the same degree of acetylation of 0.26 and differing only in average molecular weights. Hansen's Solubility Parameter Calculation. To further reveal and provide an in-depth understanding of the structure− composition−property relationships and to show the effect of the upgrading on the interactions and affinity between the different constituents of each system, the Hansen's solubility parameters (HSP) concept was used. The HSP concept is a well-established theoretical and computational approach that has been applied for many decades to study the interactions, miscibility, and compatibility of polymeric materials in different systems including wood,26 barriers,28 and more recently softwood hydrolysate formulations.16 The HSP theory models component miscibility through their nonpolar (δD), dipole− dipole (δP), and hydrogen-bonding (δH) interactions inscribed in a three-dimensional space, as illustrated in Figure 4.
films with a light brownish color. If the films are used as an internal layer in multilayer packaging films, this color will however not exert a visible effect on the resulting product. Furthermore, due to the lignin hydrophobic character, WH0based films had a lower moisture uptake at 50% RH compared to the upgraded wood hydrolysate (Supporting Information, Figure S3). This difference in moisture uptake will affect the mechanical and barrier performance of WH-based films and coatings because water acts as plasticizer. Unexpectedly, the ash persisted in upgraded WHs. This could be explained by the ionic interactions between these inorganic constituents and the hemicelluloses. A contrasting trend was observed for the LMW fractions. The ash, arabinose, lignin, rhamnose, galactose, and glucose contents were remarkably higher in LMW fractions than in the untreated and upgraded WHs, suggesting that most of the pectin was degraded to small oligo- and monosaccharides. Hydrolysate Structures. The effect of the upgrading pretreatment on the structure of hemicelluloses in the WHs was determined by evaluating the O-acetyl-4-O-methylglucuronoxylan degree of branching and degree of acetylation, calculated based on the 1H NMR spectra presented in the Figure 3 as the molar ratios of MeGlcpA to xylose and the molar ratio of acetyl to carbohydrate, respectively, assuming that all the acetyl groups are linked only to the hemicellulose.25 Hemicelluloses in wood hydrolysates WH0−WH3 are more acetylated than in the LMW fractions. In addition, the hemicelluloses in WH3 after ethanol precipitation had slightly lower DSacetyl (0.2) than in WH2 (0.26) and WH1 (0.26). However, the xylans in WH1 and WH2 had a degree of substitution similar to the untreated one. The degree of branching of xylan in the different WHs changes after upgrading pretreatments. The hemicelluloses in WH0 had a degree of branching of 0.45. The upgrading pretreatments with ultrafiltration, diafiltration, and ethanol precipitation resulted in hemicelluloses with a degree of branching of 0.37, 0.27, and 0.29, respectively. Therefore, WHs hemicellulose is composed of highly branched xylan. This suggests that the distribution of the MeGlcpA on the xylan chain is clustered27 and that the unbranched xylan was adsorbed on the surface of the solidphase cellulose during the hydrothermal treatment, thus, not dissolving in WH process water. As for the degree of acetylation, the ethanol precipitation is the upgrading route that most significantly changed the original composition of the hydrolysate. Hydrolysates Molecular Weight. The average molecular weights and polydispersity index (PDI) for commercial birch wood xylan, the untreated and upgraded WHs and their corresponding LMW fractions are shown in Table 1. The SEC traces were monomodal as shown in the Supporting Information, Figure S1. As expected, the upgrading pretreatments enriched the fraction of higher molecular weights and decreased the PDI value of WHs (assuming that the chemical composition and molecular structure of different WHs did not have a considerable influence on hydrodynamic volumes as assessed by SEC). The highest average molecular weights and lowest PDI were found in the WH3 obtained by ethanol precipitation. More extensive membrane filtration give the same increasing enrichment as expected. The average molecular weights and the PDI values of the LMWs fractions were quite similar regardless of the upgrading pretreatment. In summary, these results indicate that WH chemical structure, composition, and average molecular weight change
Figure 4. Hansen’s solubility parameter concept models the mutual affinity of components A and B through their cohesive energy parameters and Ra forming a tolerance boundary sphere in the Hansen’s space.
As mentioned earlier, the hemicellulose structure, that is, the degrees of branching and acetylation, as well as the chemical composition, varies for each WH. The δD, δP, and δH parameters of each constituent were estimated using a group contribution calculation, taking into account this variation in the structure and chemical composition for each system. The calculated HSP are then fitted into eqs 3 and 4, as previously mentioned in the Experimental Section. The calculated HSP for the hemicellulose component of each WH (herein denoted WH0-X, WH1-X, WH2-X, and WH3-X respectively) are summarized in Tables 2 and 3 together with the HSPs of highly purified BrX, and each WH with all components included (WH0−WH3). The obtained values show that the hydrogen-bonding parameter δH is higher than the nonpolar δD and dipole−dipole δP parameters for all cases, and this trend was even more pronounced for the highly purified BrX, which is reasonable given the large amount of hydroxyl groups present in the structure. This indicates that the 471
dx.doi.org/10.1021/bm201518d | Biomacromolecules 2012, 13, 466−473
Biomacromolecules
Article
Table 2. Hansen's Solubility Parameters for the WH0−WH3 Hemicelluloses and Commercial Birch Xylan BrX BrX WH0-X WH1-X WH2-X WH3-X
δD (MPa1/2)
δP (MPa1/2)
δH (MPa1/2)
δ (MPa1/2)
16.94 15.62 15.72 15.88 15.87
17.80 13.81 14.01 14.31 14.52
31.29 21.06 21.74 22.73 23.09
39.78 29.64 30.27 31.21 31.56
Table 5. Distance between the CMC and WH0−WH3 and Commercial Birch Xylan BrX in Hansen's Space distance between the CMC and WH0−WH3 and commercial birch xylan in Hansen's space Ra(CMC−BrX) Ra(CMC−WH0) Ra(CMC−WH1) Ra(CMC−WH2) Ra(CMC−WH3)
Table 3. Hansen's Solubility Parameters δD, δP, δH, and δ for Commercial Birch Xylan BrX, WH0−WH3, Lignin, and CMC BrX WH0 WH1 WH2 WH3 lignin CMC
δD (MPa1/2)
δP (MPa1/2)
δH (MPa1/2)
δ (MPa1/2)
16.94 16.06 16.11 16.26 15.93 21.90 18.7
17.80 13.83 14.02 14.30 14.52 14.10 13
31.29 20.77 21.44 22.36 23.03 16.90 24
39.78 29.68 30.26 31.13 31.54 31.05 33.1
films when compared to the CMC/BrX counterpart. These findings corroborate also with the results of the previously conducted study on softwood WH,16 showing that a softwood hydrolysate had a stronger mutual affinity with CMC than with highly purified AcGGM. On the other hand, the different upgraded WHs showed small differences in the Ra(CMC−WH) values, which did not reflect the loss in the oxygen barrier performance of the WH after upgrading pretreatments. This could be explained by the fact that the oxygen barrier performance of the coatings do depend also on several other factors such as the density and orientation of molecules which was not taken into account by the HSP approach.
hydrogen bonding interactions contribute strongly to the network of interactions that exist between the different WH components. The estimated distances between lignin and xylan from the different WH and from BrX in the Hansen's space are shown in Table 4. The radius of interaction sphere of lignin in Hansen's
■
CONCLUSION Films and coatings were successfully prepared from a hardwood-derived process water fraction, a noncellulosic polysaccharide rich wood hydrolysate (WH). The WH was recovered through several alternative pathways and mixed with CMC by water casting. The WH in the most crude state produce films with the best performance in terms of oxygen permeability. The WH was derived from the process water generated in the hydrothermal treatment of birch and aspen wood and then upgraded by either of three alternative routes: ultrafiltration, ultrafiltration + diafiltration, or ethanol precipitation. The upgrading pretreatments in general resulted in an enrichment of a higher molecular weight polysaccharide fraction, mainly consisting of hemicelluloses of the xylan type. As a result of upgrading, the xylan degree of branching was lowered, while the degree of acetylation did not change to any considerable extent. Yet, different pretreatments resulted in different product profiles with respect to structure and composition, and the most significant change was obtained by ethanol precipitation. These differences in structure and composition of the upgraded WHs had a major impact on their performance when they were used in the preparation of films and coatings. Ethanol precipitated WH, having the highest average molecular weight, the lowest lignin content, and the lowest hemicellulose degree of branching among the produced WHs, resulted in films and coatings with the highest mechanical strength but, noteworthy, with the poorest oxygen barrier function. Instead, the least upgraded WH fraction proved to be the best oxygen barrier. Furthermore, the addition of a diafiltration step contributes to an enrichment of xylan and higher molecular weights but does not result in a better oxygen barrier performance that would justify the additional cost and time of performing this step. Theoretical calculations of the constituents’ cohesive energy and mutual affinities based on the Hansen's solubility parameter approach showed that the WHs in general had a stronger mutual affinity with the CMC cocomponent than the highly purified commercial birch xylan. Stronger component affinities
Table 4. Distance between the Lignin and Xylan from the Different Wood Hydrolysates and Commercial Birch Xylan BrX in Hansen's Space distance between the lignin and xylan in Hansen's space Ra(lignin−BrX) Ra(lignin−WH0X) Ra(lignin−WH1X) Ra(lignin−WH2X) Ra(lignin−WH3X)
9.41 6.24 5.87 5.31 5.82
17.86 13.23 13.27 13.38 13.56
solubility parameter space, R0(lignin) is 13.7. Hence, Ra(lignin− xylan)/R0 > 1 for BrX, which means that it is outside the interaction sphere of lignin in the HSP space. For WH xylans, on the other hand, the ratio Ra(lignin−xylan)/R0 < 1, meaning that WH xylan is inside the sphere of interaction, showing that the lignin has a higher mutual affinity to WH xylan than to highly purified BrX. These findings reflect the important role of MeGlcpA and acetyl side groups in reducing the xylan hydrophilicity and making the WH constituents more compatible. The values of Ra(lignin−xylan) of the different WHs do not differ much; nevertheless, xylan in the WH0 has the smallest Ra(lignin−xylan), mainly due to the fact that it has the highest degree of branching. As presented in Table 5, the distance between CMC and WHs Ra(CMC−WH) in the Hansen's space is much shorter than the distance between the CMC and highly purified BrX Ra(CMC−BrX), which indicates that the WHs are more compatible with the CMC cocomponent than with commercial birch xylan. Therefore, the CMC/WH system forms generally more compact films compared to the CMC/BrX system, which is reflected through the generally higher tensile strength and Emodulus and the lower OP and OTR values of the CMC/WH 472
dx.doi.org/10.1021/bm201518d | Biomacromolecules 2012, 13, 466−473
Biomacromolecules
Article
(21) ASTM D 3985-02: Standard test method for oxygen gas transmission rate through plastic film and sheeting using a coulometric sensor; ASTM International, West Conshohocken, PA, 2002. (22) Krochta, J. M.; DeMulderJohnston, C. Food Technol. 1997, 51, 61−74. (23) Hansen, N. M. L.; Plackett, D. Biomacromolecules 2008, 9, 1493−1505. (24) Kato, Y.; Kaminaga, J.; Matsuo, R.; Isogai, A. J. Polym. Environ. 2005, 13, 261−266. (25) Fengel, D.; Wegener, G. Wood: Chemistry, Ultrastructure, Reactions; de Gruyter: Berlin, 1984. (26) Hansen, C. M.; Bjorkman, A. Holzforschung 1998, 52, 335−344. (27) Toman, R.; Kohn, R.; Malovikova, A.; Rosik, J. Collect. Czech. Chem. Commun. 1981, 46, 1405−1412. (28) Hansen, C. M. Hansen Solubility Parameters: A User’s Handbook; CRC Press: Boca Raton, FL, 2007.
contribute to forming a more homogeneous material with more immobilized chain segments in a dense disorder structure, which explains the higher mechanical and barrier performances of the WH-based films and coatings.
■
ASSOCIATED CONTENT
S Supporting Information *
Size exclusion chromatography (SEC) chromatograms of wood hydrolysates WH1−WH3, SEM images of coatings based on wood hydrolysate and carboxymethyl cellulose (WH-CMC), and moisture uptake curves at 50% relative humidity (RH) of WH-CMC-based films. This material is available free of charge via the Internet at http://pubs.acs.org.
■
ACKNOWLEDGMENTS The authors gratefully acknowledge Tetra Pak Packaging Solutions AB for financial support. Dr. Margaretha Söderqvist-Lindblad at Södra Innovation AB is thanked for kindly providing the process water.
■
REFERENCES
(1) Ragauskas, A. J.; Williams, C. K.; Davison, B. H.; Britovsek, G.; Cairney, J.; Eckert, C. A.; Frederick, W. J.; Hallett, J. P.; Leak, D. J.; Liotta, C. L.; Mielenz, J. R.; Murphy, R.; Templer, R.; Tschaplinski, T. Science 2006, 311, 484−489. (2) Thornton, J.; Ekman, R.; Holmbom, B.; Orsa, F. J. Wood Chem. Technol. 1994, 14, 159−175. (3) Persson, T.; Matusiak, M.; Zacchi, G.; Jonsson, A. S. Desalination 2006, 199, 411−412. (4) Li, H.; Saeed, A.; Jahan, M. S.; Ni, Y.; van Heiningen, A. J. Wood Chem. Technol. 2010, 30, 48−60. (5) Hartman, J.; Albertsson, A. C.; Soderqvist-Lindblad, M. J. Appl. Polym. Sci. 2006, 100, 2985−2991. (6) Albertsson, A. C.; Voepel, J.; Edlund, U.; Dahlman, O.; Soderqvist-Lindblad, M. Biomacromolecules 2010, 11, 1406−1411. (7) Persson, T.; Nordin, A. K.; Zacchi, G.; Jönsson, A. S. Appl. Biochem. Biotechnol. 2007, 137−140, 741−752. (8) Willfö r, S.; Sundberg, K.; Tenkanen, M.; Holmbom, B. Carbohydr. Polym. 2008, 72, 197−210. (9) Lawoko, M.; Henriksson, G.; Gellerstedt, G. Biomacromolecules 2005, 6, 3467−3473. (10) Lindblad, M. S.; Ranucci, E.; Albertsson, A. C. Macromol. Rapid Commun. 2001, 22, 962−967. (11) Edlund, U.; Albertsson, A. C. J. Bioact. Compat. Polym. 2008, 23, 171−186. (12) Mikkonen, K. S.; Heikkinen, S.; Soovre, A.; Peura, M.; Serimaa, R.; Talja, R. A.; Helen, H.; Hyvoenen, L.; Tenkanen, M. J. Appl. Polym. Sci. 2009, 114, 457−466. (13) Mikkonen, K. S.; Mathew, A. P.; Pirkkalainen, K.; Serimaa, R.; Xu, C. L.; Willfor, S.; Oksman, K.; Tenkanen, M. Cellulose 2010, 17, 69−81. (14) Fonseca Silva, T. C.; Habibi, Y.; Colodette, J. L.; Lucia, L. A. Soft Matter 2011, 7, 1090−1099. (15) Edlund, U.; Ryberg, Y. Z.; Albertsson, A.-C. Biomacromolecules 2010, 11, 2532−2538. (16) Ryberg, Y. Z.; Edlund, U.; Albertsson, A. C. Biomacromolecules 2011, 12, 1355−1362. (17) Dahlman, O.; Edlund, U.; Albertsson, A.-C.; SoderqvistLindblad, M. Utilization of a wood hydrolysate. Patent WO2009068525-A1, 2009. (18) Sluiter, A.; Hames, B.; Ruiz, R.; Scarlata, C.; Sluiter, J.; Templeton, D.; Crocker, D. Laboratory Analytical Procedures, National Renewable Energy Laboratory: Golden, CO, 2004. (19) Teleman, A.; Tenkanen, M.; Jacobs, A.; Dahlman, O. Carbohydr. Res. 2002, 337, 373−377. (20) ASTM D638: Standard test method for tensile properties of plastics; ASTM International, West Conshohocken, PA, 2008. 473
dx.doi.org/10.1021/bm201518d | Biomacromolecules 2012, 13, 466−473